Preliminary approach for the detection of olive trees infected by Xylella fastidiosa using a field robot and proximal sensing
Derechos de accesoopenAccess
MetadatosMostrar el registro completo del ítem
AutorLópez, Santiago; Cubero, Sergio; Aleixos, Nuria; Alegre, Vicente; Rey, Beatriz; Aguilar, Enrique; Blasco, José
Cita bibliográficaLópez, S., Cubero, S., Aleixos, N., Alegre, V., Rey, B., Aguilar, E. & Blasco, J. (2018). Preliminary approach for the detection of olive trees infected by Xylella fastidiosa using a field robot and proximal sensing. Proceedings of the European Conference on Agricultural Engineering (AgEng2018), 286-290.
A small field robot was designed and built within the framework of the H2020 project Xylella fastidiosa Active Containment Through a Multidisciplinary-Oriented Research Strategy (XF-ACTORS). The robot is remotely driven and provided with different proximal sensing equipment for the early detection of Xf in olive groves, including thermal, colour and multispectral cameras, and a 2D laser scanner (LiDAR) to obtain the 3D structure of the crop. The equipment is completed by a GPS to geolocate the data obtained and an IMU (inertial measurement unit) to correct the data captured by the LiDAR. An industrial computer triggers the sensors and controls the data acquisition, which is synchronised with the advance of the robot by means of a pulse encoder coupled to the axis of the motor. Then, crop maps can be created off-line after the analysis of the collected data to show graphically potential Xf infection in the trees. Owing to the height of the olive trees inspected, the cameras were placed on a platform that can be elevated up to 200 cm. Two batteries power the electric motors attached to the wheels, thereby allowing a continuous inspection for approximately six hours (a field of about 4 ha). A series of tests have been carried out in an olive orchard showing slight symptoms of Xf infection in the region of Apulia, southern Italy. During the first tests, the robot inspected each row in both directions with the cameras pointing to one side, so as to inspect all sides of the trees. The tests were mainly focused on the development of the mechanics, navigation systems, sensors and data acquisition. Synchronised and geolocated images of the whole crop were also captured with the cameras in different climatic conditions, as well as with the laser scanner for later comparison to the in-situ observations